For more than fifteen years, the smartphone has been the center of personal computing. It replaced cameras, music players, maps, notebooks, wallets, alarm clocks and, for many people, even desktop computers. Yet the question has never gone away: what comes after the smartphone?
The answer has seemed obvious at different moments. Smartwatches would free us from screens. VR headsets would turn computing into an immersive world. AR glasses would overlay digital information onto daily life. Earbuds would become invisible computers. But none of these categories has truly replaced the smartphone. They have become companions, not successors.
Now artificial intelligence has reopened the debate. If computers can see, listen, remember, reason and act on our behalf, perhaps the next personal device will not be a better screen. Perhaps it will be a new kind of interface entirely.
That is why OpenAI’s move into hardware with Jony Ive matters. Ive, the designer most associated with Apple’s iMac, iPod, iPhone and Apple Watch era, is now helping OpenAI imagine a new generation of AI-first personal devices. The project is still secretive, but the ambition is clear: build something that makes AI feel less like an app and more like a natural extension of the person using it.
Why the Smartphone Has Been So Hard to Replace
The smartphone is not just a device. It is an ecosystem.
It has a large touchscreen, always-on internet, cameras, sensors, payments, apps, identity, notifications and social networks. It sits in the pocket, works all day, and already has billions of users trained to interact with it. Any successor must beat the smartphone not on one feature, but on convenience, trust, battery life, privacy, cost and habit.
That is why previous contenders have struggled.
Smartwatches are useful for health, notifications and quick actions, but they are too small for deep computing. VR headsets can be powerful, but they are expensive, isolating and physically demanding. Smart glasses are improving quickly, but they still face design, battery, privacy and social-acceptance challenges. Voice assistants promised ambient computing, but for years they remained too limited and unreliable.
The smartphone survived because every “replacement” still needed it nearby.
The New Theory: AI-Native Devices
The AI-native device is different from earlier wearables because it is not trying to shrink the smartphone. It is trying to change the interaction model.
Instead of opening apps, tapping menus and typing commands, the user speaks, looks, gestures or simply allows the device to understand context. The AI becomes the interface. The hardware becomes the sensor layer.
An AI-native device may not need a large screen if it can listen, see, summarize, remember and act. It may not need hundreds of apps if agents can book, search, message, translate, buy, schedule and organize across services. It may not even look like a phone.
This is the core opportunity OpenAI and Jony Ive appear to be chasing: not a smartphone clone, but a new “personal computer” built around multimodal AI.
What Might OpenAI and Jony Ive Build?
The strongest rumors suggest that OpenAI’s first device may be small, pocketable, screen-light or screen-free, and designed to be contextually aware. It may not be a headset or traditional wearable. That points to several possible form factors.
1. A Screenless AI Companion
This could be a small object carried in a pocket, placed on a desk or clipped occasionally to clothing. It might use microphones, cameras and sensors to understand the user’s environment, then respond through voice, earbuds or a connected display.
Its advantage would be simplicity. No app grid. No doomscrolling. No visual overload. The risk is that without a screen, users may not trust what the AI is doing. Screens are not just for content; they are also for confirmation, correction and control.
2. An AI Pendant or Pin
Humane’s AI Pin showed both the appeal and danger of this category. A small wearable computer sounds futuristic, but it must solve heat, battery life, latency, privacy and usefulness immediately. If it is slower than a phone or worse than earbuds, users will abandon it.
Still, the pendant/pin idea is not dead. A better version could work if it is lighter, cheaper, more private and deeply integrated with a strong AI assistant.
3. Smart Glasses Without Heavy AR
Smart glasses may be the most practical near-term successor category. Meta’s Ray-Ban smart glasses have shown that people may accept camera-and-audio glasses if they look like normal eyewear. The most successful versions are not full AR headsets; they are stylish glasses with cameras, speakers, microphones and AI.
This matters because the winning device may arrive gradually. First, glasses capture photos, answer questions and translate speech. Then they add displays. Then they become memory and navigation tools. Eventually, they may replace many phone interactions.
4. Earbuds as the Invisible Computer
Earbuds are already socially accepted, always near the body and perfect for voice interaction. The limitation is that they cannot see. But paired with a phone, glasses or pocket camera, AI earbuds could become a powerful interface.
A future OpenAI device might even be less about one gadget and more about a personal AI system spread across earbuds, a pocket device and optional displays.
5. A New Kind of AI Phone
There are also rumors around AI-first smartphones. This may sound less revolutionary, but it could be commercially realistic. A phone remains the easiest hardware category for consumers to understand. An AI-native phone could redesign the operating system around agents instead of apps.
The problem is differentiation. Apple, Google, Samsung and Chinese smartphone makers are all adding AI features. A new AI phone would need something far stronger than “ChatGPT with a camera.” It would need a new operating model.
Lessons From Humane, Rabbit and Meta
The first wave of AI gadgets delivered a clear message: AI hardware cannot survive on novelty.
Humane’s AI Pin became a warning sign. It promised a post-smartphone future but was criticized for price, performance, heat, battery life and unclear daily utility. Rabbit’s R1 generated excitement with the idea of an AI action device, but it also faced skepticism over whether it needed separate hardware at all.
Meta’s Ray-Ban glasses, by contrast, point toward a better path. They work because they are useful even before they become revolutionary. They are sunglasses first, camera glasses second, audio devices third and AI devices fourth. That order matters. Consumers do not want to wear a science project. They want a normal object that quietly becomes more capable.
The lesson for OpenAI is simple: the next personal device must be useful on day one, socially acceptable on day two, and magical only after it has earned trust.
Why AI Changes the Hardware Equation
AI gives hardware startups a new opening because it reduces dependence on traditional app interfaces. A small device can become powerful if it has access to strong models, personal context and agentic tools.
The key capabilities are:
Context: the device understands what the user is doing, seeing or asking.
Memory: it remembers preferences, conversations, places and tasks with permission.
Agency: it can take action, not just answer questions.
Multimodality: it can combine voice, vision, text, location and personal data.
Low friction: it is faster than pulling out a phone.
This creates room for new hardware categories. A device does not need to replace every smartphone function. It only needs to own a few high-frequency moments: remembering, translating, navigating, capturing, summarizing, messaging, shopping, scheduling or assisting at work.
Opportunities for Hardware Startups
The search for the smartphone successor creates several startup opportunities.
1. Vertical AI Wearables
Instead of building a general-purpose consumer device, startups can target specific users: doctors, warehouse workers, field engineers, teachers, visually impaired users, cyclists, creators or enterprise teams. A focused product can succeed before a mass-market device is ready.
2. Smart Glasses Software Layers
The hardware may be controlled by Meta, Google, Samsung or Apple, but startups can build the agent layer: meeting assistants, memory tools, accessibility apps, real-time coaching, industrial workflows or personal productivity agents.
3. Privacy-First Personal AI
As AI devices become always-on, privacy becomes the biggest barrier. Startups that build local processing, private memory, permission controls, encrypted personal data stores or visible consent systems could become essential infrastructure.
4. AI Companions for Work
A wearable that records meetings, identifies tasks, updates CRM systems, creates calendar events and summarizes site visits could be more valuable in business than in the consumer market. Enterprises may pay for devices that save time and reduce documentation work.
5. New Input Methods
The smartphone’s touchscreen was its breakthrough. The next device needs its own input breakthrough: voice, gaze, gesture, neural wristbands, spatial pointing or context-aware automation. Startups working on natural input may become acquisition targets.
6. Accessory Ecosystems
If OpenAI, Meta or Apple creates a major AI device category, accessories will follow: charging cases, prescription lenses, privacy covers, mounts, enterprise docks, secure storage, developer kits and fashion partnerships.
The Barriers Are Still Huge
The post-smartphone dream faces serious obstacles.
Battery life remains difficult for always-on AI. Cameras and microphones raise privacy concerns. Cloud AI creates latency and cost problems. On-device AI is improving but still limited. Regulators may scrutinize always-listening and always-seeing devices. Consumers may reject devices that feel creepy, distracting or socially awkward.
There is also the app problem. Smartphones won because developers created millions of use cases. AI-native devices need an equivalent ecosystem, but not necessarily an app store. They may need agent stores, tool marketplaces, workflow libraries or trusted integrations.
Finally, there is the trust problem. A phone shows users what is happening. An AI agent may act invisibly. For a new device to succeed, it must make AI actions transparent, reversible and controllable.
Will the Smartphone Be Replaced?
Probably not quickly.
The more likely future is that the smartphone becomes less central rather than disappearing. Smart glasses may handle quick questions, translation, photos and navigation. Earbuds may handle conversation and reminders. Watches may monitor health. AI agents may complete tasks in the background. The phone may remain the secure hub for identity, payments, settings and high-attention work.
In other words, the successor to the smartphone may not be one device. It may be a personal AI network.
That is why OpenAI’s hardware effort is so important. The company is not merely trying to sell a gadget. It is trying to define how people live with AI when AI is no longer trapped inside a chat window.
Jony Ive’s design challenge is enormous. The next device must be powerful but calm, intelligent but trustworthy, personal but private, useful but not addictive. The smartphone trained us to look down. The next great device may succeed only if it helps us look back up.
The search for the smartphone’s successor is no longer about screens. It is about presence. Whoever builds the first AI device that feels natural, respectful and genuinely useful may define the next era of personal computing.